import numpy as np # ndarry 다차원 배열 (vector1차원, matrix2차원, tensor= 3차원)
import scipy # 과학계산용 패키지 ( optimization, interpolation ( 보간법) )
import sklearn.metrics.pairwise 
# accuracy_score, score 
# norm은 원점에서의 거리값
a_64 = np.array([61.22, 71.60, -65.755], dtype= np.float64)
b_64 = np.array([61.22, 71.608, -65.72], dtype=np.float64)
a_32 = a_64.astype(np.float32)
b_32 = b_64.astype(np.float32)
dist_64_np = np.array([np.linalg.norm(a_64 - b_64)], dtype=np.float64) # 차이벡터 
dist_32_np = np.array([np.linalg.norm(a_32 - b_32)], dtype=np.float32)

dist_64_sklearn = sklearn.metrics.pairwise.pairwise_distances([a_64],[b_32], metric = "cityblock")
dist_64_sklearn = sklearn.metrics.pairwise.pairwise_distances([a_64],[b_64])

from sklearn.metrics.pairwise import euclidean_distances
print("euclidean_distances", euclidean_distances([a_64],[b_64]))

euclidean_distances [[0.03590265]]

원래의 지도학습을 비지도학습하는것 ( 결과를 알기때문에 이해하기 쉬워서)

from sklearn import cluster, datasets 
import matplotlib.pyplot as plt
import seaborn as sns
%matplotlib inline
iris = datasets.load_iris() # 4개, 종속 1
X = iris.data[:, :2] # 전체데이터 150개 ( 3개종 1종당 50개의 데이터를 관찰 ) 150x2
y_iris = iris.target 
km2 = cluster.KMeans(n_clusters=2).fit(X)
km3 = cluster.KMeans(n_clusters=3).fit(X)
km4 = cluster.KMeans(n_clusters=4).fit(X)
plt.figure(figsize=(9,3))
plt.subplot(131) # 1행3열에 1번째
plt.scatter(X[:,0], X[:, 1], c=km2.labels_) # 모델
# 작아야 응집도 최대화되는
plt.title("K=2, J=%.2f" % km2.inertia_) # 응집도  ( 중심에서의 거리의 합)
plt.subplot(132)
plt.scatter(X[:,0], X[:,1], c=km3.labels_)
plt.title("K=3, J=%.2f" % km3.inertia_)
plt.subplot(133)
plt.scatter(X[:,0], X[:, 1], c=km4.labels_)
plt.title("K=4, J=%.2f" % km4.inertia_)

km4.n_iter_
km4.cluster_centers_

array([[5.90980392, 2.73529412],
       [6.85348837, 3.1       ],
       [4.76923077, 2.93076923],
       [5.21333333, 3.65333333]])

km2.labels_

array([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 0, 1, 0, 1, 0, 1, 1, 1, 0, 0, 1, 0,
       1, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, 1, 0, 0, 0,
       1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 1, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0,
       0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1])

km4.n_iter_ # iteration

6

km4.cluster_centers_ # 그룹의 특성을 이해

array([[5.90980392, 2.73529412],
       [6.85348837, 3.1       ],
       [4.76923077, 2.93076923],
       [5.21333333, 3.65333333]])

# clustering 이후에 3차원으로 시각화

from mpl_toolkits.mplot3d import Axes3D
from sklearn.cluster import KMeans
from sklearn import datasets
np.random.seed(5)
centers = [[1,1], [-1, -1], [1,-1]]
iris = datasets.load_iris()
X = iris.data
y = iris.target
estimators = {'_3': KMeans(n_clusters=3),
             '_8': KMeans(n_clusters=8)}
fignum = 1
for name, est in estimators.items():
    fig = plt.figure(fignum)
    plt.clf() # clear entire current figure 초기화
    # elev = 카메라 상하각, azim = 좌우각
    ax = Axes3D(fig, rect=[0, 0, .95, 1], elev=48, azim= 134)
    plt.cla() # 축값을 지움
    est.fit(X) # 그룹학습
    labels = est.labels_
    # 0: Sepal length, 1: Sepal width, 2: Petal length, 3: Petal width
    ax.scatter(X[:, 3], X[:, 0], X[:, 2],
              c = labels.astype(np.float), s= 100)
    ax.w_xaxis.set_ticklabels([]) # 레이블 X
    ax.w_yaxis.set_ticklabels([])
    ax.w_zaxis.set_ticklabels([])
    ax.set_xlabel('Petal width')
    ax.set_ylabel('Sepal length')
    ax.set_zlabel('Petal length')
    plt.title(name)
    fignum = fignum + 1
plt.show()

<ipython-input-19-b97f5529ebd5>:20: DeprecationWarning: `np.float` is a deprecated alias for the builtin `float`. To silence this warning, use `float` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.float64` here.
Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
  c = labels.astype(np.float), s= 100)
<ipython-input-19-b97f5529ebd5>:20: DeprecationWarning: `np.float` is a deprecated alias for the builtin `float`. To silence this warning, use `float` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.float64` here.
Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
  c = labels.astype(np.float), s= 100)

from sklearn.datasets import load_digits
from sklearn.preprocessing import scale
# kmeans와 scale의 관계 : 중요 - 사이즈에 따라 거리값이 불공평
import matplotlib.pyplot as plt
digits = load_digits()
data = scale( digits. data) # Z점수 표준화
def print_digits(images, labels):
    f =plt.figure(figsize=(10,2))
    plt.subplots_adjust(top=1, bottom=0, hspace=0, wspace=0.05)
    i = 0
    while (i < 10 and i < images.shape[0]): # 10장까지만, 장수 10장이 안되면 
        ax = f.add_subplot(1,10, i+1) # 이미지 출력은 1부터
        ax.imshow(images[i], cmap= plt.cm.bone)
        ax.grid(False)
        ax.table
        ax.set_title(labels[i])
        ax.xaxis.set_ticks([])
        ax.yaxis.set_ticks([])
        plt.tight_layout()
        i += 1
print_digits(digits.images, range(10))

이미지를 저장

• 데이터로 다루거나, 이미지로 다루거나

digits.data.shape # 1797장

(1797, 64)

digits.images.shape # 이미지로 찍을 때 차원

(1797, 8, 8)

digits.images[1]

array([[ 0.,  0.,  0., 12., 13.,  5.,  0.,  0.],
       [ 0.,  0.,  0., 11., 16.,  9.,  0.,  0.],
       [ 0.,  0.,  3., 15., 16.,  6.,  0.,  0.],
       [ 0.,  7., 15., 16., 16.,  2.,  0.,  0.],
       [ 0.,  0.,  1., 16., 16.,  3.,  0.,  0.],
       [ 0.,  0.,  1., 16., 16.,  6.,  0.,  0.],
       [ 0.,  0.,  1., 16., 16.,  6.,  0.,  0.],
       [ 0.,  0.,  0., 11., 16., 10.,  0.,  0.]])

from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test, images_train, images_test \
= train_test_split(
    data, digits.target, digits.images, test_size=0.25, random_state=42)

# 클러스터링을 하면 라벨이 결정 => 지도학습한 것과 일치
# 중심 10개 생성 -> 새로운 데이터 들어 오면
from sklearn.cluster import KMeans
# 0~9 : 분류
clf = KMeans(init="k-means++", n_clusters=10, random_state=42)
clf.fit(X_train)
print_digits(images_train, clf.labels_)

y_pred = clf.predict(X_test)
def print_cluster(images, y_pred, cluster_number):
    images = images[y_pred == cluster_number]
    y_pred = y_pred[y_pred == cluster_number]
    print_digits(images, y_pred) # 클러스터당 10장씩 출력
for i in range(10):
    print_cluster(images_test, y_pred, i )

<Figure size 720x144 with 0 Axes>

from sklearn.datasets.samples_generator import make_blobs
X, y =make_blobs(n_samples=100, centers=5, random_state=101)

C:\Users\1\anaconda3\lib\site-packages\sklearn\utils\deprecation.py:143: FutureWarning: The sklearn.datasets.samples_generator module is  deprecated in version 0.22 and will be removed in version 0.24. The corresponding classes / functions should instead be imported from sklearn.datasets. Anything that cannot be imported from sklearn.datasets is now part of the private API.
  warnings.warn(message, FutureWarning)

import seaborn as sns
from sklearn.cluster import KMeans
elbow = []
for i in range(1, 20):
    kmeans = KMeans(n_clusters = i, init = 'k-means++',
                   random_state=101)
    kmeans.fit(X)
    elbow.append(kmeans.inertia_)
sns.lineplot(range(1,20), elbow,color='blue')
plt.rcParams.update({'figure.figsize':(10, 7.5), 'figure.dpi': 100})
plt.title('elbow method')
plt.show()

C:\Users\1\anaconda3\lib\site-packages\seaborn\_decorators.py:36: FutureWarning: Pass the following variables as keyword args: x, y. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.
  warnings.warn(

from imageio import imread
import matplotlib.pyplot as plt
img = imread("garden.jpg")
plt.imshow(img)
img.shape
x,y,z = img.shape
long_img = img.reshape(x*y, z)
long_img.shape

(446400, 3)

from sklearn.datasets import load_sample_image # 이미지 학습
from sklearn.utils import shuffle
n_colors = 16
china = load_sample_image("china.jpg") # 0~255구성
china = np.array(china, dtype = np.float64)/ 255 # 이미지 정규화 
w, h ,d = original_shape = tuple(china.shape)
# 컬러이미지는 기본적으로 3차원이다. (8,8)
print(w, h, d) # 427 640 3 # 가로 , 세로 , 컬값
assert d==3 # 컬러이미지를 로딩했는지 확인 
image_array = np.reshape(china, (w * h, d)) # 2차원으로 직렬화
print("Kmeans 데이터 피팅")
# 427  * 640
image_array_sample = shuffle(image_array, random_state=0)[:1000]
kmeans = KMeans(n_clusters=n_colors, random_state=0).fit(image_array_sample)
# 결과 중심 : 64
# 라벨링 : 427  * 640
print("(k-means) 예측")
labels = kmeans.predict(image_array) # 427  * 640

# 압축, 이미지를 64Color로 재표현
def recreate_image(codebook, labels, w, h): # 중심값팔레트, 중심값인덱스, 이미지크기
    d = codebook.shape[1] # 2차원
    image= np.zeros((w,h,d)) # 이미지 사이즈만큼 초기화
    label_idx = 0
    for i in range(w):
        for j in range(h):
            image[i][j] = codebook[labels[label_idx]]
            label_idx += 1
    return image

427 640 3
Kmeans 데이터 피팅
(k-means) 예측

plt.figure(1)
plt.clf()
ax = plt.axes([0,0,1,1])
plt.axis('off')
plt.title('Original 이미지 (96,615 colors)')
plt.imshow(china)
plt.figure(2)
plt.clf()
ax = plt.axes([0,0,1,1])
plt.axis('off')
plt.title('Vector Quantized(VQ) image (64 colors, K-Menas)')
plt.imshow(recreate_image(kmeans.cluster_centers_, labels, w, h))

<matplotlib.image.AxesImage at 0x22fb9ef8ac0>

C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:238: RuntimeWarning: Glyph 51060 missing from current font.
  font.set_text(s, 0.0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:238: RuntimeWarning: Glyph 48120 missing from current font.
  font.set_text(s, 0.0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:238: RuntimeWarning: Glyph 51648 missing from current font.
  font.set_text(s, 0.0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:201: RuntimeWarning: Glyph 51060 missing from current font.
  font.set_text(s, 0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:201: RuntimeWarning: Glyph 48120 missing from current font.
  font.set_text(s, 0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:201: RuntimeWarning: Glyph 51648 missing from current font.
  font.set_text(s, 0, flags=flags)

문제:

• 웹에서 이미지를 다운받으시오
• 이미지 사이즈를 확인하시오
• 컬러를 16컬러로 벡터 양자화하시오
• 양자화된 값으로 이미지를 표현해 보시오
• 원래이미지와 양자화된 이미지의 사이즈를 비교해 보시오

%matplotlib inline
import matplotlib.pyplot as plt
from sklearn import cluster
image = plt.imread("서현진1.jpg")
plt.figure(figsize=(15,8))
plt.imshow(image)

<matplotlib.image.AxesImage at 0x22fb9f60e20>

from sklearn.datasets import load_sample_image # 이미지 학습
from sklearn.utils import shuffle
n_colors = 16
sh = imread("서현진1.jpg") # 0~255구성
sh = np.array(sh, dtype = np.float64)/ 255 # 이미지 정규화 
w, h ,d = original_shape = tuple(sh.shape)
# 컬러이미지는 기본적으로 3차원이다. (8,8)
print(w, h, d) # 427 640 3 # 가로 , 세로 , 컬값
assert d==3 # 컬러이미지를 로딩했는지 확인 
image_array = np.reshape(sh, (w * h, d)) # 2차원으로 직렬화
print("Kmeans 데이터 피팅")
# 427  * 640
image_array_sample = shuffle(image_array, random_state=0)[:1000]
kmeans = KMeans(n_clusters=n_colors, random_state=0).fit(image_array_sample)
# 결과 중심 : 64
# 라벨링 : 427  * 640
print("(k-means) 예측")
labels = kmeans.predict(image_array) # 427  * 640

# 압축, 이미지를 64Color로 재표현
def recreate_image(codebook, labels, w, h): # 중심값팔레트, 중심값인덱스, 이미지크기
    d = codebook.shape[1] # 2차원
    image= np.zeros((w,h,d)) # 이미지 사이즈만큼 초기화
    label_idx = 0
    for i in range(w):
        for j in range(h):
            image[i][j] = codebook[labels[label_idx]]
            label_idx += 1
    return image

656 656 3
Kmeans 데이터 피팅
(k-means) 예측

plt.figure(1)
plt.clf()
ax = plt.axes([0,0,1,1])
plt.axis('off')
plt.title('Original 이미지 (96,615 colors)')
plt.imshow(image)
plt.figure(2)
plt.clf()
ax = plt.axes([0,0,1,1])
plt.axis('off')
plt.title('Vector Quantized(VQ) image (16 colors, K-Menas)')
plt.imshow(recreate_image(kmeans.cluster_centers_, labels, w, h))

<matplotlib.image.AxesImage at 0x22fb9beeac0>

C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:238: RuntimeWarning: Glyph 51060 missing from current font.
  font.set_text(s, 0.0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:238: RuntimeWarning: Glyph 48120 missing from current font.
  font.set_text(s, 0.0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:238: RuntimeWarning: Glyph 51648 missing from current font.
  font.set_text(s, 0.0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:201: RuntimeWarning: Glyph 51060 missing from current font.
  font.set_text(s, 0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:201: RuntimeWarning: Glyph 48120 missing from current font.
  font.set_text(s, 0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:201: RuntimeWarning: Glyph 51648 missing from current font.
  font.set_text(s, 0, flags=flags)

print(image.shape)
# kilo byte
print((image.shape[0] * image.shape[1] * image.shape[2])/1024, "kb")

(656, 656, 3)
1260.75 kb

x, y, z = image.shape
image = np.array(image, dtype = np.float64)/ 255
image_2d =image.reshape(x * y, z) # 컬러값들로만 표현
print(image_2d.shape)

(430336, 3)

kmeans_cluster =cluster.KMeans(n_clusters=16)
kmeans_cluster.fit(image_2d)
cluster_centers =kmeans_cluster.cluster_centers_
cluster_centers

array([[0.81521686, 0.65172505, 0.58347005],
       [0.23335047, 0.13015071, 0.09727977],
       [0.76118724, 0.59018885, 0.52382254],
       [0.54119204, 0.37641395, 0.31796993],
       [0.3817041 , 0.23126174, 0.18524984],
       [0.93008311, 0.82807302, 0.78674661],
       [0.12102418, 0.05318226, 0.02485256],
       [0.86349607, 0.71565464, 0.64408491],
       [0.70027473, 0.51969257, 0.45952031],
       [0.18012572, 0.09035189, 0.05857787],
       [0.45966165, 0.30348683, 0.25063392],
       [0.62878855, 0.44828408, 0.39002871],
       [0.30145963, 0.17775191, 0.13915939],
       [0.89519583, 0.77041429, 0.71744371],
       [0.64828172, 0.26867463, 0.26539219],
       [0.78356185, 0.70167092, 0.6949171 ]])

cluster_labels = kmeans_cluster.labels_
print(cluster_labels) # 픽셀 개수 만큼

[15 15 15 ...  9  9  6]

print(len(cluster_labels))
print("압축사이즈=", ((cluster_centers.shape[0]*
                cluster_centers.shape[1]) + len(cluster_labels))/ 1024)

plt.figure(figsize=(15,8))
plt.imshow(cluster_centers[cluster_labels].reshape(x,y,z))

430336
압축사이즈= 420.296875

<matplotlib.image.AxesImage at 0x22fba21c490>

import matplotlib
from matplotlib import font_manager, rc
import platform

if platform.system() == 'Windows':
# 윈도우인 경우
    font_name = font_manager.FontProperties(fname="c:/Windows/Fonts/malgun.ttf").get_name()
    rc('font', family=font_name)
else:    
# Mac 인 경우
    rc('font', family='AppleGothic')
    
matplotlib.rcParams['axes.unicode_minus'] = False

import numpy as np
import matplotlib.pyplot as plt

from sklearn.datasets import make_blobs
X, label = make_blobs(100, centers =1 )
from sklearn.cluster import KMeans
kmeans = KMeans(n_clusters=1)
kmeans.fit(X)
f, ax = plt.subplots(figsize=(7,5))
ax.set_title('원형이상치제거')
ax.scatter(X[:,0], X[:,1], label='Points')
ax.scatter(kmeans.cluster_centers_[:,0],
          kmeans.cluster_centers_[:,1], label='Centroid', color='r')
ax.legend(loc='best')

<matplotlib.legend.Legend at 0x2a0bef53d90>

C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:238: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0.0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:201: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0, flags=flags)

import matplotlib
matplotlib.font_manager._rebuild()

matplotlib.__version__

'3.3.2'

distances = kmeans.transform(X) # 데이터를 ceter point와의 거리값변경
sorted_idx =np.argsort(distances.ravel())[::-1][:5]

데이터, 인덱스 10, 8, 2, 인덱스 
0 Z 일  2,1,4,3,0
1 B 이
2 A
3 F
4 C

f, ax =plt.subplots(figsize=(7,5))
ax.set_title('Single Cluster')
ax.scatter(X[:,0], X[:, 1], label='Points')
ax.scatter(kmeans.cluster_centers_[:, 0],
          kmeans.cluster_centers_[:,1],
          label='Centroid', color='r')
ax.scatter(X[sorted_idx][:,0],
          X[sorted_idx][:,1],
          label='이상치', edgecolors='g',
          facecolors='none', s=100)
ax.legend(loc='best')
new_X = np.delete(X, sorted_idx, axis=0)
new_kmeans = KMeans(n_clusters=1)
new_kmeans.fit(new_X)

KMeans(n_clusters=1)

C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:238: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0.0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:201: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0, flags=flags)

문제 : 이상치가 중심점 좌표에 영향을 미치는 것을 확인하고 이상치 제거전과 후의 중심점을 그래프로 출력해 보시오

f, ax =plt.subplots(figsize=(7,5))
ax.set_title('이상치 제거후')
ax.scatter(new_X[:,0], new_X[:, 1], label='Pruned Points')
ax.scatter(kmeans.cluster_centers_[:, 0],
          kmeans.cluster_centers_[:,1],
          label='이전중심', color='r', s=80, alpha=.5)
ax.scatter(new_kmeans.cluster_centers_[:,0],
          new_kmeans.cluster_centers_[:,1],
          label='이후중심', color='m', s=80, alpha=.5)
ax.legend(loc='best')

<matplotlib.legend.Legend at 0x2a0c44cea90>

import mglearn # 모델 시각화를 위해 생성모델
import matplotlib.pyplot as plt
from sklearn.cluster import AgglomerativeClustering # 계층적 : 병합
from sklearn.datasets import make_blobs
X, y = make_blobs(random_state=1)

agg = AgglomerativeClustering(n_clusters=3)
assignment = agg.fit_predict(X) # 그룹번호를 할당
# 모델 출력을 위해서
mglearn.discrete_scatter(X[:,0],X[:,1], assignment)
# loc :location
plt.legend(["클러스터 0", "클러스터 1", "클러스터 2"], loc="best")
plt.xlabel("특성 0")
plt.ylabel("특성 1")

Text(0, 0.5, '특성 1')

C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:238: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0.0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:201: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0, flags=flags)

from sklearn.preprocessing import StandardScaler
from sklearn.datasets import make_moons
import matplotlib.pyplot as plt
from sklearn.cluster import DBSCAN
import mglearn
X, y = make_moons(n_samples=200, noise=0.05, random_state=0)
scaler = StandardScaler()
scaler.fit(X)
# clustering 에서도 정규화하면 좋다.
X_scaled = scaler.transform(X)

dbscan = DBSCAN() # scikits : eps, min_samples
clusters = dbscan.fit_predict(X_scaled)

plt.scatter(X_scaled[:, 0], X_scaled[:, 1], c=clusters,
           cmap=mglearn.cm2, s=60, edgecolors='black')
plt.xlabel("특성 0")
plt.ylabel("특성 1")

Text(0, 0.5, '특성 1')

C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:238: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0.0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:201: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0, flags=flags)

mglearn.discrete_scatter(X_scaled[:,0], X_scaled[:,1],clusters)

[<matplotlib.lines.Line2D at 0x2a0c53768b0>,
 <matplotlib.lines.Line2D at 0x2a0c5376bb0>]

C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:238: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0.0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:201: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0, flags=flags)

mglearn.plots.plot_dbscan()

min_samples: 2 eps: 1.000000  cluster: [-1  0  0 -1  0 -1  1  1  0  1 -1 -1]
min_samples: 2 eps: 1.500000  cluster: [0 1 1 1 1 0 2 2 1 2 2 0]
min_samples: 2 eps: 2.000000  cluster: [0 1 1 1 1 0 0 0 1 0 0 0]
min_samples: 2 eps: 3.000000  cluster: [0 0 0 0 0 0 0 0 0 0 0 0]
min_samples: 3 eps: 1.000000  cluster: [-1  0  0 -1  0 -1  1  1  0  1 -1 -1]
min_samples: 3 eps: 1.500000  cluster: [0 1 1 1 1 0 2 2 1 2 2 0]
min_samples: 3 eps: 2.000000  cluster: [0 1 1 1 1 0 0 0 1 0 0 0]
min_samples: 3 eps: 3.000000  cluster: [0 0 0 0 0 0 0 0 0 0 0 0]
min_samples: 5 eps: 1.000000  cluster: [-1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1]
min_samples: 5 eps: 1.500000  cluster: [-1  0  0  0  0 -1 -1 -1  0 -1 -1 -1]
min_samples: 5 eps: 2.000000  cluster: [-1  0  0  0  0 -1 -1 -1  0 -1 -1 -1]
min_samples: 5 eps: 3.000000  cluster: [0 0 0 0 0 0 0 0 0 0 0 0]

문제

다음 데이터에 대하여 eps 와 min_samples를 조절하여 3개의 그룹으로

• clustering하도록 조절해 보시오
• 그리고 결과를 시각화해 보시오

import numpy as np

from sklearn.cluster import DBSCAN
from sklearn import metrics
from sklearn.datasets import make_blobs
from sklearn.preprocessing import StandardScaler


# #############################################################################
# Generate sample data
centers = [[1, 1], [-1, -1], [1, -1]]
X, labels_true = make_blobs(n_samples=750, centers=centers, cluster_std=0.4,
                            random_state=0)

X = StandardScaler().fit_transform(X)

# #############################################################################
# Compute DBSCAN
db = DBSCAN(eps=0.3, min_samples=10).fit(X)
core_samples_mask = np.zeros_like(db.labels_, dtype=bool)
core_samples_mask[db.core_sample_indices_] = True
labels = db.labels_

# Number of clusters in labels, ignoring noise if present.
n_clusters_ = len(set(labels)) - (1 if -1 in labels else 0)
n_noise_ = list(labels).count(-1)

print('Estimated number of clusters: %d' % n_clusters_)
print('Estimated number of noise points: %d' % n_noise_)
print("Homogeneity: %0.3f" % metrics.homogeneity_score(labels_true, labels))
print("Completeness: %0.3f" % metrics.completeness_score(labels_true, labels))
print("V-measure: %0.3f" % metrics.v_measure_score(labels_true, labels))
print("Adjusted Rand Index: %0.3f"
      % metrics.adjusted_rand_score(labels_true, labels))
print("Adjusted Mutual Information: %0.3f"
      % metrics.adjusted_mutual_info_score(labels_true, labels))
print("Silhouette Coefficient: %0.3f"
      % metrics.silhouette_score(X, labels))

# #############################################################################
# Plot result
import matplotlib.pyplot as plt

# Black removed and is used for noise instead.
unique_labels = set(labels)
colors = [plt.cm.Spectral(each)
          for each in np.linspace(0, 1, len(unique_labels))]
for k, col in zip(unique_labels, colors):
    if k == -1:
        # Black used for noise.
        col = [0, 0, 0, 1]

    class_member_mask = (labels == k)

    xy = X[class_member_mask & core_samples_mask]
    plt.plot(xy[:, 0], xy[:, 1], 'o', markerfacecolor=tuple(col),
             markeredgecolor='k', markersize=14)

    xy = X[class_member_mask & ~core_samples_mask]
    plt.plot(xy[:, 0], xy[:, 1], 'o', markerfacecolor=tuple(col),
             markeredgecolor='k', markersize=6)

plt.title('Estimated number of clusters: %d' % n_clusters_)
plt.show()

Estimated number of clusters: 3
Estimated number of noise points: 18
Homogeneity: 0.953
Completeness: 0.883
V-measure: 0.917
Adjusted Rand Index: 0.952
Adjusted Mutual Information: 0.916
Silhouette Coefficient: 0.626

C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:238: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0.0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:201: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0, flags=flags)

from sklearn.cluster import DBSCAN
from sklearn import metrics
from sklearn.datasets.samples_generator import make_blobs
from sklearn.preprocessing import StandardScaler
import mglearn
centers = [[1,1], [-1, -1], [1, -1]]
X , labels_true = make_blobs(n_samples=750, centers=centers,
                            cluster_std=0.4, random_state=0)
scaler =StandardScaler()
scaler.fit(X)
X_scaled = scaler.transform(X)
bscan = DBSCAN(eps=0.3, min_samples=10).fit(X)
clusters = dbscan.fit_predict(X_scaled)
plt.scatter(X_scaled[:, 0], X_scaled[:, 1], c=clusters,
           cmap=mglearn.cm2, s=60, edgecolors='black')
plt.xlabel("특성 0")
plt.ylabel("특성 1")

Text(0, 0.5, '특성 1')

C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:238: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0.0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:201: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0, flags=flags)

mglearn.discrete_scatter(X_scaled[:,0], X_scaled[:,1],clusters)

[<matplotlib.lines.Line2D at 0x2a0c5368670>]

C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:238: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0.0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:201: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0, flags=flags)

mglearn.plots.plot_dbscan()

min_samples: 2 eps: 1.000000  cluster: [-1  0  0 -1  0 -1  1  1  0  1 -1 -1]
min_samples: 2 eps: 1.500000  cluster: [0 1 1 1 1 0 2 2 1 2 2 0]
min_samples: 2 eps: 2.000000  cluster: [0 1 1 1 1 0 0 0 1 0 0 0]
min_samples: 2 eps: 3.000000  cluster: [0 0 0 0 0 0 0 0 0 0 0 0]
min_samples: 3 eps: 1.000000  cluster: [-1  0  0 -1  0 -1  1  1  0  1 -1 -1]
min_samples: 3 eps: 1.500000  cluster: [0 1 1 1 1 0 2 2 1 2 2 0]
min_samples: 3 eps: 2.000000  cluster: [0 1 1 1 1 0 0 0 1 0 0 0]
min_samples: 3 eps: 3.000000  cluster: [0 0 0 0 0 0 0 0 0 0 0 0]
min_samples: 5 eps: 1.000000  cluster: [-1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1]
min_samples: 5 eps: 1.500000  cluster: [-1  0  0  0  0 -1 -1 -1  0 -1 -1 -1]
min_samples: 5 eps: 2.000000  cluster: [-1  0  0  0  0 -1 -1 -1  0 -1 -1 -1]
min_samples: 5 eps: 3.000000  cluster: [0 0 0 0 0 0 0 0 0 0 0 0]

팀별 과제

• tsclus를 학습하고 time series를 clustering하는 예제를 구현해 보시오
• tslearn = kmeans + DTW(dynamic time wrapping) + timeseries를 결합한 패키지

Import Modules

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

from tslearn.clustering import TimeSeriesKMeans

C:\Users\1\anaconda3\lib\site-packages\tslearn\clustering\kmeans.py:16: UserWarning: Scikit-learn <0.24 will be deprecated in a future release of tslearn
  warnings.warn(

Load Data

df = pd.read_csv('ex_eample.csv', index_col=0)
df.head()

	t0	t1	t2	t3	t4	t5	t6
idx
0	12	27	30	48	59	60	76
1	11	30	31	44	58	66	70
2	16	25	36	50	59	68	76
3	19	26	38	47	56	64	73
4	18	22	30	43	58	61	72

plt.plot(df.T)
plt.show()

Preprocessing

avg_list = []
std_list = []

for idx, row in df.iterrows():
    avg_list.append(row.mean())
    std_list.append(row.std())

df['avg'] = avg_list
df['std'] = std_list
df.head()

	t0	t1	t2	t3	t4	t5	t6	avg	std
idx
0	12	27	30	48	59	60	76	44.571429	22.463727
1	11	30	31	44	58	66	70	44.285714	21.638810
2	16	25	36	50	59	68	76	47.142857	22.363874
3	19	26	38	47	56	64	73	46.142857	19.777332
4	18	22	30	43	58	61	72	43.428571	20.911378

for i in range(len(df)):
    for j in range(0, 7):
        df.iloc[i, j] = (df.iloc[i, j] - df.iloc[i, 7]) / df.iloc[i, 8]
df.head()

	t0	t1	t2	t3	t4	t5	t6	avg	std
idx
0	-1.449957	-0.782213	-0.648665	0.152627	0.642305	0.686822	1.399081	44.571429	22.463727
1	-1.538241	-0.660189	-0.613976	-0.013204	0.633782	1.003488	1.188341	44.285714	21.638810
2	-1.392552	-0.990117	-0.498253	0.127757	0.530192	0.932627	1.290346	47.142857	22.363874
3	-1.372423	-1.018482	-0.411727	0.043340	0.498406	0.902910	1.357976	46.142857	19.777332
4	-1.216016	-1.024733	-0.642166	-0.020495	0.696818	0.840281	1.366310	43.428571	20.911378

df = df.drop(['avg', 'std'], axis=1)
df.head()

	t0	t1	t2	t3	t4	t5	t6
idx
0	-1.449957	-0.782213	-0.648665	0.152627	0.642305	0.686822	1.399081
1	-1.538241	-0.660189	-0.613976	-0.013204	0.633782	1.003488	1.188341
2	-1.392552	-0.990117	-0.498253	0.127757	0.530192	0.932627	1.290346
3	-1.372423	-1.018482	-0.411727	0.043340	0.498406	0.902910	1.357976
4	-1.216016	-1.024733	-0.642166	-0.020495	0.696818	0.840281	1.366310

Model1: Euclidean

model = TimeSeriesKMeans(n_clusters=3, metric='euclidean', random_state=121)

result = model.fit_predict(df)
result

array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 2, 2, 1, 1, 2, 1, 1, 1, 2, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1], dtype=int64)

df['clst'] = result
df.head()

	t0	t1	t2	t3	t4	t5	t6	clst
idx
0	-1.449957	-0.782213	-0.648665	0.152627	0.642305	0.686822	1.399081	0
1	-1.538241	-0.660189	-0.613976	-0.013204	0.633782	1.003488	1.188341	0
2	-1.392552	-0.990117	-0.498253	0.127757	0.530192	0.932627	1.290346	0
3	-1.372423	-1.018482	-0.411727	0.043340	0.498406	0.902910	1.357976	0
4	-1.216016	-1.024733	-0.642166	-0.020495	0.696818	0.840281	1.366310	0

clst0 = df[df['clst']==0]
clst0.drop(['clst'], axis=1, inplace=True)
plt.figure(figsize=(10, 4))
plt.plot(clst0.T)
plt.show()

C:\Users\1\anaconda3\lib\site-packages\pandas\core\frame.py:4308: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  return super().drop(
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:238: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0.0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:201: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0, flags=flags)

clst1 = df[df['clst']==1]
clst1.drop(['clst'], axis=1, inplace=True)
plt.figure(figsize=(10, 4))
plt.plot(clst1.T)
plt.show()

C:\Users\1\anaconda3\lib\site-packages\pandas\core\frame.py:4308: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  return super().drop(
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:238: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0.0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:201: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0, flags=flags)

clst2 = df[df['clst']==2]
clst2.drop(['clst'], axis=1, inplace=True)
plt.figure(figsize=(10, 4))
plt.plot(clst2.T)
plt.show()

C:\Users\1\anaconda3\lib\site-packages\pandas\core\frame.py:4308: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  return super().drop(
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:238: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0.0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:201: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0, flags=flags)

Model2: DTW

model2 = TimeSeriesKMeans(n_clusters=3, metric='dtw', random_state=121)

result2 = model2.fit_predict(df)
result2

array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 1, 1, 1, 2, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1], dtype=int64)

df['clst'] = result2
df.head()

	t0	t1	t2	t3	t4	t5	t6	clst
idx
0	-1.449957	-0.782213	-0.648665	0.152627	0.642305	0.686822	1.399081	0
1	-1.538241	-0.660189	-0.613976	-0.013204	0.633782	1.003488	1.188341	0
2	-1.392552	-0.990117	-0.498253	0.127757	0.530192	0.932627	1.290346	0
3	-1.372423	-1.018482	-0.411727	0.043340	0.498406	0.902910	1.357976	0
4	-1.216016	-1.024733	-0.642166	-0.020495	0.696818	0.840281	1.366310	0

clst0 = df[df['clst']==0]
clst0.drop(['clst'], axis=1, inplace=True)
plt.figure(figsize=(10, 4))
plt.plot(clst0.T)
plt.show()

C:\Users\1\anaconda3\lib\site-packages\pandas\core\frame.py:4308: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  return super().drop(
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:238: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0.0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:201: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0, flags=flags)

clst1 = df[df['clst']==1]
clst1.drop(['clst'], axis=1, inplace=True)
plt.figure(figsize=(10, 4))
plt.plot(clst1.T)
plt.show()

C:\Users\1\anaconda3\lib\site-packages\pandas\core\frame.py:4308: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  return super().drop(
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:238: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0.0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:201: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0, flags=flags)

clst2 = df[df['clst']==2]
clst2.drop(['clst'], axis=1, inplace=True)
plt.figure(figsize=(10, 4))
plt.plot(clst2.T)
plt.show()

C:\Users\1\anaconda3\lib\site-packages\pandas\core\frame.py:4308: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  return super().drop(
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:238: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0.0, flags=flags)
C:\Users\1\anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py:201: RuntimeWarning: Glyph 8722 missing from current font.
  font.set_text(s, 0, flags=flags)